The APOBEC3 cytidine deaminases generate two of the most prevalent mutational signatures in human cancer, found across a majority of tumour types. Which family member drives that mutagenesis, and what governs it, were long-standing questions.
Using genetic deletion across breast, bladder, lung and lymphoma models, we established that APOBEC3A is the dominant source. Deleting it removes the signatures. Deleting the far more abundant APOBEC3B does not, in most contexts. That was counterintuitive, since APOBEC3A is barely detectable in bulk measurements of tumours, and it left a paradox. A scarcely observed enzyme was dominating the mutational landscape of epithelial cancers.
The resolution is that the enzyme does not work steadily. Mutations arrive in bursts, concentrated in a minority of branches of a cell lineage rather than spread evenly across it. Single-cell sequencing showed APOBEC3A confined to a small fraction of cells, and a reporter knocked into the endogenous locus let us capture them. The high state is transient, resolving within days rather than marking a stable subpopulation.
Those cells turn out to occupy a squamous differentiation state, appearing as a plastic, stress-associated programme in cancers that are not squamous by lineage. A squamous transcription factor gates entry, and the enzyme is not a passive passenger. Its own activity reinforces part of the programme, stabilising the state in which it acts. Cancer therapy engages the same circuit, which raises the possibility that treatment accelerates mutagenesis in the cells that survive it.
We are now defining what controls entry into that state, and measuring what a single passage through it costs the genome.
Read more at maciejowskilab.org.
